第 59 章 有意识和无意识的心理过程障碍
尽管认知神经科学作为一门重要的新学科出现于 20 世纪末，但术语认知的确切含义仍然难以捉摸。该术语
在不同的上下文中以不同的方式使用。在一个极端，认知神经科学中的术语” 认知” 意味着旧术语” 信息处理” 的
含义。从这个意义上说，认知就是大脑所做的事情。当认知神经科学家说视觉特征或运动行为由神经活动表示
时，他们使用的是来自信息处理的概念。从这个角度来看，认知语言在神经活动和行为的描述之间架起了一座
桥梁，因为相同的术语可以应用于这 2 个领域。
在另一个极端，术语” 认知” 指的是那些对意识体验的形成至关重要的更高层次的过程。这就是认知疗法的
含义，这是一种由亚伦 · 贝克和阿尔伯特 · 埃利斯开创并从行为疗法发展而来的治疗方法。认知疗法不是试图直
接改变患者的行为，而是旨在改变患者的态度和信念（文本框 59.1）。
文本框 59.1 (认知疗法)
♠
对基于弗洛伊德无意识动机理论心理治疗的不满在 20 世纪中叶加剧。这些理论不仅与实验心理学无
关，而且没有经验证据表明心理动力学治疗确实有效。
实验室研究中出现的第一种替代性心理疗法被称为行为疗法。这种方法的基本假设是，不适应的行
为是学习的，因此可以通过应用巴甫洛夫和斯金纳的刺激反应学习原理来消除。例如，一个被狗袭击的孩
子可能会对所有的狗感到恐惧，但如果孩子知道条件刺激（狗的视觉）之后没有非条件刺激（被咬），这
种恐惧反应可能会消失。
行为疗法被证明对恐惧症快速有效，但许多精神障碍的特点是思维不适应，而不是行为不适应。20
世纪 60 年代，亚伦 · 贝克和阿尔伯特 · 埃利斯开创了一种新的治疗方法，利用学习原理来改变思想而不
是行为。这被称为认知疗法或认知行为疗法。
这种形式的治疗在抑郁症的治疗中特别成功。抑郁症通常与消极的想法（例如，一个人只记得发生在
他/她身上的坏事）和消极的态度（例如，认为他/她永远不会实现自己的目标）有关。认知治疗师教他们
的客户如何减少消极想法的频率，并将他们的消极态度转变为积极态度。
在通常的说法中，术语认知意味着思考和推理，一种更接近其拉丁词根了解或感知的用法。因此，《牛津英
语词典》将其定义为“认识的行动或能力”。事实上，我们通过对感官的原始数据进行思考和推理来了解世界。
这个想法隐含在我们对多种认知障碍的描述中。脑损伤后，一些患者无法再处理感官提供的输入。这种类
型的障碍首先由弗洛伊德描述，他称之为失认症或知识丧失（第 59 章）。失认症可以有多种形式。视觉失认症
患者可以看得很清楚，但不再能够识别或理解他所看到的东西。患有面部失认症的患者在识别面孔方面存在特
殊问题。患有听觉失认症的患者可能听得很好，但无法识别口语。
认知有时从出生就受损，因此一个人很难获得知识。这可能会导致一般智力低下，或者，如果问题更局限，
会导致特定的学习困难，例如阅读障碍（难以学习书面语言）或孤独症（难以了解其他思想）。最后，认知会变
得功能失调，以至于获得的关于世界的知识是错误的。这些思维障碍导致与精神分裂症等主要精神疾病相关的
错误观念（幻觉）和错误信念（妄想）。
59.1 有意识和无意识的认知过程具有不同的神经相关性
认知（通过思考和推理获得知识）是意识的 3 个组成部分之一（见第 42 章对情绪的意识方面的讨论，通常
称为感受）。另外
2
个是情感和意志。过去人们理所当然地认为，思考和推理是在有意识的自主控制下进行的，
没有意识就不可能有认知。然而，到 19 世纪末，弗洛伊德发展了一种无意识心理过程理论，并提出许多人类行
为是由我们不知道的内部过程引导。
对神经科学更为直接和重要的是无意识推理的概念，它最初由亥姆霍兹提出。亥姆霍兹是第一个进行定量
心理物理学实验并测量周围神经传入信号传导速度的人。在这些实验之前，人们假设感觉信号会立即（以光速）
59.2 在脑损伤后可以以夸张的形式看到感知过程中的有意识和无意识过程之间的差异
到达大脑，但亥姆霍兹表明神经传导实际上非常缓慢。他还指出，反应时间甚至更慢。这些观察表明，大脑在处
理感官刺激和我们对物体的意识感知之间进行了大量的工作。亥姆霍兹得出结论，大脑中发生的大部分事情都
不是有意识的，进入意识的东西（即被感知的东西）取决于无意识的推论。换句话说，大脑使用来自感官的证据
来决定最有可能引起感觉器官活动的物体身份，但这是在我们没有意识到的情况下进行的。
这种观点在亥姆霍兹同时代人中极为不受欢迎，事实上，今天仍然如此。大多数人认为，意识是进行推理所
必需的，道德责任需由有意识推理的决策来承担。如果可以在没有意识的情况下进行推论，那么就没有赞扬或
责备的伦理基础。亥姆霍兹关于无意识推理的观点在很大程度上被忽视了。
尽管如此，到 20 世纪中叶，越来越多的证据支持大多数认知过程从未进入意识的观点。随着电子计算机的
发展和人工智能研究的出现，研究人员开始研究机器如何以及在多大程度上感知超越自身的世界。很快就会变
得明显，许多乍一看很简单的感知过程在定义为一组计算时实际上非常复杂。
视觉感知是最好的例子。在 1960 年代，几乎没有人意识到制造能够识别物体形状和外观的机器有多么困难，
因为这对我们来说似乎太容易了。我看着窗外，看到建筑物、树木、花朵和人。我并不了解这种感知背后的任何
心理过程；我对所有这些目标的意识似乎是瞬间和直接的。事实证明，在包含许多重叠目标的典型杂乱视觉场
景中，教机器如何计算出哪些边缘与哪个对象对应是非常困难的。视觉的计算方法揭示了我们对世界看似毫不
费力的感知所依赖的潜在神经过程。类似的过程是所有感官知觉的基础，尤其是将声音感知为语音的基础。大
多数神经科学家现在认为，我们没有意识到认知过程，只有我们的知觉。
无意识认知过程的证据不仅来自人工智能研究，还来自脑损伤患者的认知研究。无意识过程对行为的影响
可以在某些“盲视”患者身上得到最显著的证明，这是
劳伦斯
·
魏斯克朗茨
在
1970
年代首次描述的一种疾病。这
些患者的初级视觉皮层有病变，并声称在受损区域的视野部分看不到任何东西。然而，当被要求猜测时，他们能
够检测到简单的视觉属性，例如运动或颜色，这比偶然预期的要好得多。尽管在视野的盲区对物体没有基于感
官的感知，但这些患者确实拥有关于物体的无意识信息，并且这些信息可用于指导他们的行为。
另一个例子是右顶叶病变引起的单侧忽视（第
17 章）。患有这种疾病的患者视力正常，但他们似乎无法察
觉到前方空间左侧的物体。有些患者甚至忽略了左侧的个别物体。如图 59.1.1 所示，在约翰 · 马歇尔和彼得 · 哈
里根的一项实验中，向患者展示了两幅房屋的图纸。一所房子的左侧着火了。当被问及房子之间是否有任何差
异时，患者回答“没有”。但当被问及他们更愿意住在哪所房子时，他们选择了没有着火的房子。因此，这种选
择是基于意识中未表示的信息做出的。盲视和单侧忽视只是丰富的无意识认知过程存在的 2 个例子，我们无法
通过内省获得这些证据。
目前，神经科学最令人兴奋的研究领域之一是寻找由弗朗西斯 · 克里克和克里斯托弗 · 科赫发起的意识的神
经相关物。目的是证明与有意识和无意识认知过程相关的神经活动之间的定性差异。这项研究很重要，不仅因
为它可以为我们解答意识功能这一难题，还因为它与我们对许多神经和精神疾病的理解有关。患有某些认知障
碍患者的怪异经历和妄想信念曾一度被认为无法理解而被忽视。认知神经科学为我们提供了一个框架来理解这
些经验和信念如何从正常认知机制的特定改变中产生。
59.2 在脑损伤后可以以夸张的形式看到感知过程中的有意识和无意识过程之
间的差异
感官刺激和知觉之间的关系远非直接的。如图 59.2.1 所示，知觉可以在感觉刺激没有任何变化的情况下发
生变化，如花瓶与人脸交变图和内克尔立方体等模棱两可的图所示。相反，在观察者没有意识到这种变化的情
况下，感官刺激可能会发生巨大变化（感知保持不变）。一个令人信服的例子是变化盲。
为了证明变化盲，我们构建了 1 个复杂场景的 2 个版本。在罗恩 · 伦辛克开发的一个著名示例中，图片由一
架停在机场跑道上的军用运输机组成。在这 2 个版本之一中，缺少引擎。如果这两张图片在计算机屏幕上交替
显示，但严重地穿插着空白屏幕，哪怕明确指出差异，参与者可能需要几分钟才能注意到复杂场景 2 个版本之
间的差异（另一个示例参见图 25.2.1）。
鉴于这些现象，我们可以探索当感觉刺激没有变化时与知觉变化相关的神经活动。同样，我们可以发现感
1305
59.2 在脑损伤后可以以夸张的形式看到感知过程中的有意识和无意识过程之间的差异
Chapter 59 / Disorders of Conscious and Unconscious Mental Processes 1475
It rapidly became clear that many perceptual processes
that at first seem simple are actually very complex
when defined as a set of computations.
Visual perception is the prime example. In the
1960s, almost no one realized how difficult it would
be to build machines that could recognize the shape
and appearance of objects, because it seems so easy
for us. I look out of the window and I see buildings,
trees, flowers, and people. I am not aware of any men-
tal processes behind this perception; my awareness of
all these objects seems instantaneous and direct. It
turns out that teaching a machine how to work out
which edges go with which object in a typical clut-
tered visual scene containing many overlapping
objects is exceptionally difficult. The computational
approach to vision revealed the underlying neural
processes on which our seemingly effortless percep-
tion of the world depends. Similar processes underlie
all sensory perception and especially the perception
of sounds as speech. Most neuroscientists now believe
that we are not conscious of cognitive processes, only
our perceptions.
The evidence for unconscious cognitive processes
comes not only from artificial intelligence studies but
also studies of cognition in people with brain damage.
The effects of unconscious processes on behavior can
be demonstrated most strikingly in certain patients
with “blind sight,” a disorder first delineated in the
1970s by Lawrence Weiskrantz. These patients have
lesions in the primary visual cortex and claim to see noth-
ing in the part of the visual field served by the damaged
area. Nevertheless, when asked to guess, they are able
to detect simple visual properties such as movement
or color far better than is expected by chance. Despite
having no sensory-based perception of objects in the
blind parts of the visual field, these patients do possess
unconscious information about the objects, and this
information is available to guide their behavior.
Another example is unilateral neglect caused by
lesions in the right parietal lobe (Chapter 17). Patients
with this disorder have normal vision, but they seem
unware of objects on the left side of the space in front
of them. Some patients even ignore the left side of
individual objects. In one experiment by John Marshall
and Peter Halligan, patients were shown two draw-
ings of a house. The left side of one house was on fire
(Figure 59–1). When asked if there were any differ-
ences between the houses, patients replied “no.” But
when asked which house they would prefer to live
in, they chose the house that was not burning. This
choice was thus made based on information that was
not represented in consciousness. Blind sight and uni-
lateral neglect are just two examples of the abundant
右顶后皮层
Figure 59–1 Unconscious processing in cases of spatial
neglect.After damage to the right parietal lobe, many patients
seem to be unaware of the left side of space (unilateral neglect
syndrome). When such patients are shown the two drawings
reproduced here, they say that the two houses look the same.
However, they also say that they would prefer to live in the
lower house, indicating that they have unconsciously processed
the image of the fire in the other house. (Adapted from Marshall
and Halligan 1988.)
empirical evidence for the existence of unconscious
cognitive processes, evidence not available to us
through introspection.
Currently, one of the most exciting areas of
investigation in neuroscience concerns the search for
the neural correlates of consciousness initiated by Fran-
cis Crick and Christopher Koch. The aim is to dem-
onstrate qualitative differences between the neural
activity associated with conscious and unconscious
cognitive processes. This research is important not only
because it may give us answers to the difficult ques-
tion of the function of consciousness but also because
it is relevant to our understanding of many neurologi-
cal and psychiatric disorders. The weird experiences
Kandel-Ch59_1473-1487.indd 1475 18/12/20 11:41 AM
图 59.1.1: 空间忽视情况下的无意识处理。右顶叶受损后，许多患者似乎意识不到左侧空间（单侧忽视综合症）。
当这些患者看到这里复制的两幅图时，他们说这两座房子看起来一样。然而，他们也表示他们更愿意住在下层
房子里，这表明他们已经不自觉地处理了另一栋房子发生火灾的形象
[533]
。
1476 Part IX / Diseases of the Nervous System
Figure 59–2 Ambiguous figures.If you stare at the figure
on the left (the Rubin figure), you sometimes see a vase and
sometimes two faces looking at each other. If you stare at
the figure on the right (the Necker cube), you see a three-
dimensional cube, but the front face of the cube is
sometimes seen at the bottom left and sometimes at the top
right. In each figure, the brain finds two equally good, but mutu-
ally exclusive, interpretations of what is there. Our conscious
perception spontaneously alternates between these two
interpretations.
and delusional beliefs of patients with certain cogni-
tive disorders were once dismissed as beyond under-
standing. Cognitive neuroscience provides us with a
framework for understanding how these experiences
and beliefs can arise from specific alterations in normal
cognitive mechanisms.
Differences Between Conscious and
Unconscious Processes in Perception Can Be
Seen in Exaggerated Form After Brain Damage
The relationship between sensory stimulation and
perception is far from direct. Perception can change
without any change in sensory stimulation, as illus-
trated by ambiguous figures such as the Rubin figure
and the Necker cube (Figure 59–2). Conversely, a big
change in sensory stimulation can occur without the
observer being aware of this change—the perception
remains constant. A compelling example of this is
change blindness.
To demonstrate change blindness, two versions of
a complex scene are constructed. In one well-known
example developed by Ron Rensink, the picture con-
sists of a military transport plane standing on an air-
port runway. In one of the two versions, an engine is
missing. If these two pictures are shown in alternation
on a computer screen, but critically interspersed with
a blank screen, it can take minutes to notice the dif-
ference even though it is immediately obvious when
pointed out. (See Figure 25-8 for another example.)
In light of these phenomena, we can explore the
neural activity associated with changes in perception
when there is no change in sensory stimulation. Like-
wise, we can discover whether changes in sensory
input are registered in the brain even if not repre-
sented in consciousness. We can ask whether there is
some qualitative difference between the neural activity
associated with conscious as opposed to unconscious
processes.
Two important results have emerged from studies
of the neural activity associated with specific types of
conscious percepts. First, certain kinds of percepts are
related to neural activity in specific areas of the brain.
Those brain areas that are specialized for recognition
of certain kinds of objects (eg, faces, words, land-
scapes) or for certain visual features (eg, color, motion)
are more active when the object or the feature is con-
sciously perceived (Figure 59–3). For example, when
we perceive the faces in the Rubin figure, there is more
activity in the area of the fusiform gyrus, which is spe-
cialized for the processing of faces.
This observation also applies to deviant percep-
tion (hallucinations). After degeneration of the periph-
eral visual system leading to blindness, some patients
experience intermittent visual hallucinations (Charles
Bonnet syndrome). These hallucinations vary from one
patient to another: Some patients see colored patches,
others see grid-like patterns, and some even see faces.
Dominic ffytche found that these hallucinations are
associated with increased activity in the secondary
visual cortex, and the content of the hallucination is
related to the specific locus of activity (Figure 59–4).
Schizophrenic patients frequently experience com-
plex auditory hallucinations, which usually have the
form of voices talking to or about the patient. These
Kandel-Ch59_1473-1487.indd 1476 18/12/20 11:41 AM
图 59.2.1: 模棱两可的数字。如果你盯着左边的人物（花瓶与人脸交变图），你有时会看到一个花瓶，有时会看到
两张面孔。如果你盯着右边的图（内克尔立方体），你会看到一个三维立方体，但立方体的正面有时会出现在左
下角，有时会出现在右上角。在每个图形中，大脑都会找到
2
个同样好的但相互排斥的解释。我们的意识感知自
发地在这 2 种解释之间交替。
1306
59.2 在脑损伤后可以以夸张的形式看到感知过程中的有意识和无意识过程之间的差异
官输入的变化是否在大脑中记录下来，即使没有在意识中表现出来。我们可以问，与有意识过程相关的神经活
动与无意识过程相关的神经活动之间是否存在一些质的差异？
对与特定类型的意识知觉相关的神经活动的研究得出了 2 个重要结果。首先，某些类型的知觉与大脑特定
区域的神经活动有关。如图 59.2.2 所示，当有意识地感知物体或特征时，那些专门用于识别某些种类的物体（例
如面部、文字、风景）或某些视觉特征（例如，颜色、运动）的大脑区域会更加活跃。例如，当我们感知花瓶与
人脸交变图中的人脸时，专门处理人脸的梭状回区域活动较多。
Chapter 59 / Disorders of Conscious and Unconscious Mental Processes 1477
Figure 59–3 Neural activity associated with ambiguous
visual information.An ambiguous stimulus was created by
simultaneously presenting a face to one eye and a house to
the other eye. Brain activity was measured while subjects
observed these images. Subjects were instructed to press a
button whenever a spontaneous switch in perception occurred
(because of binocular rivalry). When the face is perceived (left),
activity increases in the fusiform face area (FFA); when the
house is perceived (right), activity increases in the parahip-
pocampal place area (PPA). (Abbreviation: MRI, magnetic reso-
nance imaging.) (Reproduced, with permission, from Tong et al.
1998. Copyright © 1998 by Cell Press.)
梭状回
面孔区
海马旁回
梭状回
面孔区
海马旁回
房子 面孔
% 核磁共振成像信号
0.0
0.2
0.4
0.6
0.8
1.0
0–8 –4 04
81
2
4812 –8 –4
报告感知切换后的时间（秒）
房子面孔
刺激
感知
海马旁回
梭状回
面孔区
hallucinations are associated with activity in the audi-
tory cortex.
These observations suggest that conscious experi-
ence may result from activity in certain cortical regions.
This idea is difficult to test experimentally, but in the
1950s, the neurosurgeon Wilder Penfield found that
electrical stimulation of the cortex in patients undergo-
ing neurosurgery can generate a conscious experience.
More recently, it has been found that transcranial mag-
netic stimulation of the cortex in the region of V5/MT
can lead to seeing moving light flashes.
The second important conclusion drawn from stud-
ies that seek to correlate neural activity and specific
percepts is that activity in a specialized area is neces-
sary but not sufficient to yield conscious experience. For
example, in the change blindness paradigm, subjects are
often unaware of large changes in the picture they are
viewing. If the change involves a face, activity is elicited
in the fusiform gyrus whether or not the subject is aware
of the change. But when the sensory change is also per-
ceived consciously, there is, in addition, activity in the
parietal and frontal cortices (Figure 59–5).
These observations are relevant to our understand-
ing of unilateral neglect. Since objects on the left side
still elicit neural activity in the visual cortex, it may
be that the damage in the right parietal cortex simply
prevents the formation of conscious representations of
objects on the left side of space. Nevertheless, this sen-
sory activity can support an unconscious inference in
patients that they would not want to live in the house
that is burning on the left side.
Stimuli that do not enter awareness can also elicit
overt responses. A face with a fearful expression elic-
its a fear response in the autonomic nervous system,
measured as an increase in skin conductance (galvanic
response) because of sweating. This response occurs
even if the face is immediately followed by another
visual stimulus, such that the face is not consciously
perceived. There may be an advantage to having a
rapid but low-resolution system for recognizing dan-
gerous things. We jump first; only later, on the basis of
a slow, high-resolution system, are we able to identity
the object that made us jump (Chapter 48). Damage in
one or the other of these two recognition systems can
explain certain otherwise puzzling neurological and
psychiatric disorders.
Prosopagnosia is a perceptual disorder in which
faces are no longer recognizable. The patient knows
Kandel-Ch59_1473-1487.indd 1477 18/12/20 11:41 AM
图 59.2.2: 与模糊视觉信息相关的神经活动。通过同时向一只眼睛呈现一张脸和向另一只眼睛展示一所房子，产
生了一种模糊的刺激。在受试者观察这些图像的同时测量大脑活动。受试者被要求在感知发生自发切换时按下
按钮（因为双眼竞争）。当人脸被感知（左）时，梭状回面孔区的活动增加；当房子被感知（右）时，海马旁回
的活动增加
[534]
。
这种观察也适用于异常感知（幻觉）。在导致失明的周围视觉系统退化后，一些患者会出现间歇性幻视（查
尔斯 · 庞奈综合症）。这些幻觉因患者而异：一些患者看到彩色斑块，一些患者看到网格状图案，有些甚至看到
面孔。如图 59.2.3 所示，多米尼克 · 费切发现这些幻觉与二级视觉皮层的活动增加有关，幻觉的内容与特定的活
动部位有关。精神分裂症患者经常会出现复杂的幻听，通常表现为有人与患者交谈或谈论患者的声音。这些幻
觉与听觉皮层的活动有关。
这些观察结果表明，有意识的体验可能源于某些皮层区域的活动。这个想法很难通过实验检验，但在 1950
年代，神经外科医生怀尔德 · 潘菲尔德发现，对接受神经外科手术的患者的皮层进行电刺激可以产生有意识的体
验。最近，已经发现对
五级视觉皮层
/
内侧颞叶
区域的皮层进行经颅磁刺激可以导致看到移动的闪光。
从试图将神经活动与特定知觉相关联的研究中得出的第二个重要结论是，专门区域的活动是必要的，但不
足以产生有意识的体验。例如，在变化盲范式中，受试者通常不会意识到他们正在查看的图片发生了巨大变化。
如果变化涉及面部，则无论受试者是否意识到变化，梭状回都会引起活动。但是，如图 59.2.4 所示，当感觉变化
也被有意识地感知到时，顶叶和额叶皮层就会另外活动。
这些观察与我们对单侧忽视的理解有关。由于左侧的物体仍然会在视觉皮层中引起神经活动，因此右侧顶
叶皮层的损伤可能只是阻止了对空间左侧物体的有意识表征的形成。然而，这种感觉活动可以支持患者的无意
识推断，即他们不想住在左侧燃烧的房子里。
没有进入意识的刺激也可以引起明显的反应。一张带有恐惧表情的脸会在自主神经系统中引起恐惧反应，表
现为出汗导致的皮肤电导（电流反应）增加。即使面部紧随其后是另一个视觉刺激，这种反应也会发生，这样面
部就不会被有意识地感知到。拥有快速但分辨率低的系统来识别危险事物可能会有优势。我们先跳过具体的物
1307
59.2 在脑损伤后可以以夸张的形式看到感知过程中的有意识和无意识过程之间的差异
1478 Part IX / Diseases of the Nervous System
Figure 59–4 Neural activity associated with visual halluci-
nations.Some patients with damage to the retina experience
visual hallucinations. The location of the neural activity and
the content of the hallucination are related. The experience of
colors, patterns, objects, or faces is associated with heightened
activity (red) in specific regions of inferior temporal cortex. The
blue area is the fusiform gyrus. (Reproduced, with permission,
from ffytche et al. 1998. Copyright © 1998 Springer Nature.)
受试者 1
受试者 3
受试者 2
受试者 4
he is looking at a face but cannot recognize the face,
even a beloved face known for years. The problem
is specific to faces, since the patient may still be able
to recognize the person from their clothes, gait, and
voice. However, patients with prosopagnosia are able
to identify faces unconsciously. They show autonomic
responses to familiar faces and do better than chance
when asked to guess whether or not a face shown to
them belongs to a person who is familiar. In fact, their
awareness of the autonomic (emotional) responses
elicited by a face may enable them to judge familiarity.
Capgras syndrome, a delusion that is occasion-
ally observed in schizophrenic patients and in some
patients suffering from brain injury or dementia,
produces a more unsettling experience. These patients
firmly believe that someone close to them, usually a
husband or wife, has been replaced by an impostor.
They claim that the person, although similar if not
identical in appearance, is in fact someone else. Often,
this delusion is acted on with the demand that the
impostor leave the house.
Hadyn Ellis and Andy Young have suggested that
this bizarre delusion is the mirror phenomenon of
prosopagnosia. According to this view, the circuitry
for face recognition is intact, but the circuitry that
mediates the emotional response to the face is not.
As a result, patients recognize the person in front of
them but, because the emotional response is lacking,
Kandel-Ch59_1473-1487.indd 1478 18/12/20 11:41 AM
图 59.2.3: 与视觉幻觉相关的神经活动。一些视网膜受损的患者会出现幻视。神经活动的位置与幻觉的内容有关。
颜色、图案、物体或面孔的体验与下颞皮层特定区域的活动增强（红色）有关。蓝色区域是梭状回
[535]
。
1308
59.2 在脑损伤后可以以夸张的形式看到感知过程中的有意识和无意识过程之间的差异
Chapter 59 / Disorders of Conscious and Unconscious Mental Processes 1479
Figure 59–5 Brain activity with and without aware-
ness.Activity in the fusiform face area increases when the face
viewed by subjects changes, whether subjects are unaware of
the change or conscious of it. When subjects are aware of the
change, activity in parietal and frontal cortex also increases.
(Reproduced, with permission, from Beck et al. 2001.)
梭状回
A 无意识检测
B 有意识的报告
顶叶皮层 额叶
皮层
顶叶皮层
feel that there is something fundamentally wrong. This
account has been partially confirmed by the observa-
tion that these patients do not have normal autonomic
responses to familiar faces.
This explanation implies that Capgras delusions
are not the consequence of disordered thinking but of
disordered experience. A patient sees the face of his
wife without having the normal emotional response.
The conclusion that this is not his wife but an impostor
is a cognitive response to this abnormal experience, the
mind’s attempt to explain the experience.
The Control of Action Is Largely Unconscious
The sense that we are in control of our own actions is a
major component of consciousness. But are we aware
of all aspects of our own actions? David Milner and Mel
Goodale studied a patient known as D.F. who demon-
strates a striking lack of awareness of certain aspects of
her own actions. As a result of damage to her inferior
temporal lobe caused by carbon monoxide poisoning,
D.F. suffers from form agnosia—she is unable to identify
the shapes of things. She cannot distinguish a square
from an oblong card and cannot describe the orienta-
tion of a slot. Yet when she picks up the oblong card
to place it through the slot, she orients her hand and
forms her grasp appropriately because of the uncon-
scious operation of visuomotor circuits (Figure 59–6).
This sort of unconscious guidance is not unique to
patients with brain damage. It is simply revealed more
starkly in the case of D.F. because the system that nor-
mally brings visual information about shape into con-
sciousness is impaired. Indeed, we can all make rapid
and accurate grasping movements without being
aware of the perceptual and motor information that
is being used to control these movements. Sometimes,
we are not even aware of having made the movement.
This largely unconscious system for visually guided
reaching and grasping is analogous to, and probably
overlaps with, the rapid but poor-resolution system
associated with fear responses.
Although we may not be aware of the perceptual
and motor details of actions like reaching and grasping,
Kandel-Ch59_1473-1487.indd 1479 18/12/20 11:41 AM
图 59.2.4: 有意识和无意识的大脑活动。无论受试者是否意识到或意识到这种变化，当受试者看到的面部发生变
化时，梭状回面孔区的活动会增加。当受试者意识到这种变化时，顶叶和额叶皮层的活动也会增加
[536]
。
1309
59.3 行动的控制在很大程度上是无意识的
体；随后在一个缓慢的高分辨率系统的基础上，我们才能够识别出让我们跳跃的物体（第 48 章）。这 2 个识别系
统中的一个或另一个的损坏可以解释某些令人费解的神经和精神疾病。
面容失认症是一种知觉障碍，表现为无法再辨认面孔。病人知道他在看一张脸，但无法认出那张脸，即使是
一张熟悉多年的心爱的脸。这个问题是面部特有的，因为患者可能仍然能够从他们的衣服、步态和声音中认出
这个人。然而，面容失认症患者能够在不知不觉中识别面孔。他们对熟悉的面孔表现出自主反应，并且在被要求
猜测展示给他们的面孔是否属于熟悉的人时，他们比随机猜测做出更好的反应。事实上，他们对一张脸引起的
自主（情绪）反应的意识可能使他们能够判断熟悉程度。
替身综合症是一种错觉，偶尔会在精神分裂症患者和一些脑损伤或痴呆症患者身上观察到，它会产生更令
人不安的体验。这些患者坚信他们身边的某个人（通常是丈夫或妻子）已经被冒名顶替者所取代。他们声称虽
然这个人在外貌上相似，甚至一模一样，但实际上是另一个人。通常，这种妄想症会导致患者要求冒名顶替者离
开家。
海丁 · 埃利斯和安迪 · 杨认为这种奇怪的错觉是面容失认症的镜像现象。根据这种观点，面部识别的神经回
路是完整的，但对面部产生情绪反应的神经回路却并非如此。因此，患者能够认出面前的人，但由于缺乏情绪反
应，他们觉得有些地方从根本上就是错的。通过观察发现，这些患者对熟悉的面孔没有正常的自主反应，这在一
定程度上证实了上述观点。
这种解释意味着 替身错觉不是思维混乱的结果，而是体验混乱的结果。一个病人看到他妻子的脸时没有正
常的情绪反应。于是得出结论，这不是他的妻子，而是一个冒名顶替者，这是对异常体验的认知反应，是大脑试
图解释这种体验的一种尝试。
59.3 行动的控制在很大程度上是无意识的
我们对自己行为的控制感是意识的一个重要组成部分。但是我们是否意识到自己行为的所有方面？大卫 · 米
尔纳和梅尔 · 古德尔研究了一位名叫 D.F. 的患者，她展示了对自己行为某些方面极端缺乏意识的情况。由于一
氧化碳中毒导致她的下颞叶受损，D.F. 患有形状失认症（她无法识别事物的形状）。她无法区分正方形和长方形
卡片，也无法描述插槽的方向。然而，如图 59.3.1 所示，当她拿起长方形卡片将其放入槽中时，由于视觉运动回
路的无意识操作，她会适当地调整手的方向并形成抓握动作。
这种无意识的引导并不是脑损伤患者独有的。在 D.F. 案例中，它只是更加明显地显现出来，因为正常情况
下将关于形状的视觉信息带入意识的系统受损了。事实上，我们都可以做出快速准确的抓握动作，而无需意识
到用于控制这些动作的知觉和运动信息。有时，我们甚至没有意识到自己做了动作。这种用于视觉引导的伸手
和抓取的很大程度上是无意识的系统类似于与恐惧反应相关的快速但分辨率低的系统，并且可能与之重叠。
虽然我们可能没有意识到像伸手和抓握等动作的感知和运动的细节，但我们对于一些自己的行为是否处于
控制之下有着鲜明的认识，即我们能意识到我们所引起的行为与那些不由自主发生的行为之间的差异。本杰明 ·
李贝特研究了受控实验中的自主行为现象。他要求他的受试者“每当他们感到有这样做的冲动时”就举起一根
手指，并报告他们产生这种冲动的时间。他的受试者毫不费力地可靠地报告了这一主观体验的时间。与此同时，
李贝特使用脑电图来测量“准备电位”，这是一种大脑活动的变化，发生在受试者做出任何自主运动之前 1 秒内。
受试者报告感到有抬起手指冲动的时间发生在这种准备电位开始后数百毫秒。这一结果在哲学家和神经科学家
之间引起了很多关于自由意志存在的讨论。如果脑活动可以在一个人意识到自己有冲动执行某项动作之前预测
出这个动作，那么这是否意味着我们对于自由意志行为的体验是一种幻觉？
尽管李贝特的结果已被广泛复制，但他的实验方案对我们理解自由意志的相关性仍存在争议。举起一根手
指不是我们经常执行的动作。行动通常有目标。例如，我们可能会按一个按钮来响铃。当我们的行动遵循我们期
望的目标时，我们会觉得我们可以控制自己的行动。正是这种主观体验给了我们一种能动性，成为事件的起因。
将李贝特的范式应用于此类行为，帕特里克 · 哈格德发现了“故意绑定”现象。如图 59.3.2 所示，当一个有意的
动作（按下一个按钮）之后是它的预期目标（听到一个音调）时，这些事件在主观上被体验为在时间上绑定在一
起。
我们的行动与其目标之间的这种时间绑定提供了我们代理感的经验标记，因为更强的代理感与更大程度的
1310
59.3 行动的控制在很大程度上是无意识的
1480 Part IX / Diseases of the Nervous System
Figure 59–6 Action can be controlled by unconscious
stimuli.A patient, D.F., with damage to the inferior temporal
cortex, is unable to recognize objects based on their shape
(form agnosia). She cannot align the tablet with the orienta-
tion of the slot (perceptual matching) because she is not con-
sciously aware of the orientation of either the tablet or the slot.
However, when she is asked to put the tablet through the slot
in a quick movement, she orients her hand rapidly and accu-
rately. Presumably, the movement is driven by visuomotor
computations of which the subject is unaware. (Adapted, with
permission, from Milner and Goodale 1995.)
Figure 59–7 We experience our actions and their effects as
bound together in time.When subjects are asked to press a
button that triggers a sound 250 ms later, they experience their
action and the sound as occurring closer together (subjective
time) than they actually are (objective time). In contrast, when
their finger moves involuntarily through trans cranial magnetic
stimulation (TMS) of motor cortex, the movements and the
sound are experienced as further apart compared to objective
time. Temporal binding occurs only when the movement is
intended and deliberate and thus is a marker of the experience
of agency. (Based on Haggard, Clark, and Kalogeras 2002.)
将卡插入
插槽
控制
腹侧视觉流受损
的受试者
感知
方向
匹配
we are vividly aware of being in control of some of our
actions—we are aware of a difference between actions
that we cause and those that happen involuntarily.
Benjamin Libet studied the phenomenon of voluntary
action in controlled experiments. He asked his subjects
to lift a finger “whenever they felt the urge to do so”
and to report the time at which they had this urge.
His subjects had no difficulty in reliably reporting the
time of this subjective experience. At the same time,
Libet used electroencephalography to measure the
“readiness potential,” a change in brain activity that
occurs up to 1 second before a subject makes any vol-
untary movement. The time at which subjects reported
feeling the urge to lift a finger occurred hundreds
of milliseconds after the beginning of this readiness
potential. This result has generated much discussion
among philosophers as well as neuroscientists con-
cerning the existence of free will. If brain activity can
predict an action before a person is aware of having the
urge to perform that action, does this mean that our
experience of freely willing actions is an illusion?
Although Libet’s result has been widely repli-
cated, the relevance of his experimental protocol for
our understanding of free will remains controversial.
Lifting one finger is not an action that we often per-
form. Actions usually have goals. For example, we
might press a button in order to ring a bell. When our
actions are followed by the goal we expect, we feel
that we are in control of our actions. It is this subjec-
tive experience that gives us a sense of agency, of being
the cause of events. Applying Libet’s paradigm to such
actions, Patrick Haggard discovered the phenomenon
of “intentional binding.” When a deliberate movement
(pressing a button) is followed by its intended goal
(hearing a tone), these events are experienced subjec-
tively as bound together in time (Figure 59–7).
This temporal binding of our actions to their goals
provides an empirical marker of our sense of agency,
since a stronger sense of agency is associated with a
greater degree of binding. If a movement occurs pas-
sively, caused for example by magnetic stimulation to
the brain, then intentional binding is decreased; we
actually perceive the time between movement and out-
come as longer than the actual physical time.
Our sense of agency is closely linked to our belief
in free will and to the idea that people can be held
Subjective time
Objective time
Tone
Tone
ToneAction
250 ms
Voluntary
action
Involuntary
action
Action
Action
Kandel-Ch59_1473-1487.indd 1480 18/12/20 11:41 AM
图 59.3.1: 行动可以通过无意识的刺激来控制。下颞皮层受损的患者 D.F. 无法根据物体的形状识别物体（形状失
认症）。她无法将平板电脑与插槽的方向对齐（感知匹配），因为她没有意识到平板电脑或插槽的方向。然而，当
她被要求快速将平板电脑穿过插槽时，她会快速准确地调整手的方向。据推测，运动是由受试者不知道的视觉
运动计算驱动的
[537]
。
图 59.3.2: 我们体验到我们的行为和它们的影响是及时结合在一起的。当要求受试者在 250 毫秒后按下触发声音
的按钮时，他们会体验到他们的动作和声音的发生时间（主观时间）比实际情况（客观时间）更近。相比之下，
当他们的手指通过运动皮层的经颅磁刺激不由自主地移动时，与客观时间相比，运动和声音会感觉更远。时间
约束只有在运动是有意和深思熟虑时才会发生，因此是代理体验的标志
[538]
。
1311
59.4 有意识地回忆是一个创造性的过程
绑定相关联。如果运动是被动发生的，例如由对大脑的磁刺激引起，那么有意的束缚就会减少；我们实际上会感
知运动和结果之间的时间比实际的物理时间更长。
我们对代理感的感知与我们对自由意志的信仰以及人们在有意识地执行行动时能够被追究责任的观念密切
相关。当与具有道德后果的结果相关联时，故意约束力会增加。它减少了其他人命令的动作，而不是自由执行的
动作。这些结果并没有解决自由意志是否存在的问题，但它们表明我们自由行动的有意识体验在创造社会责任
规范方面起着重要作用。这些规范对于维持社会凝聚力至关重要。
在运动领域和感知领域都存在着无意识推理。我们的代理体验由 2 个部分组成：我们先前的期望和行动结
果的感官后果。如果实际感觉与我们的预期不符，我们会感到惊讶，例如当我们拿起一个比预期轻得多的物体
时（第 30 章）。然而，如果结果证实了我们的预期，我们就会很少注意实际的感官证据，我们会体验到我们预期
会发生的事情，而不是实际发生的事情。
如图 59.3.3 所示，皮埃尔 · 富尔纳雷和马克 · 珍妮罗德要求受试者使用电脑鼠标画一条垂直线。受试者看不
到他们的手，因此看不到计算机在屏幕上显示的线条。令人惊讶的结果是，受试者没有意识到他们已经将手向
左移动 10° 角以在屏幕上产生垂直线。这种意识的缺乏的现象出现了高达 15° 的偏差中。当受试者被指示不要看
屏幕而只是重复他们刚刚做过的动作时，他们并没有重现他们做过的异常动作，而是画出他们认为自己做过的
笔直向前的动作。似乎只要实现了目标（画一条直截了当的线），我们就会体验到预期的感官反馈，而不是实际
的感官反馈。
Chapter 59 / Disorders of Conscious and Unconscious Mental Processes 1481
Figure 59–8 Actions can be modified unconsciously.Subjects
are asked to draw a straight line with a computer mouse. They can
see the line on the screen but not their hand movement. The com-
puter is programmed to systematically distort the line displayed
on the screen. In the result shown here, the subject had to move
his hand 10° to the left to produce a vertical line on the screen.
Subjects are not aware of making such adjustments. (Adapted,
with permission, from Fourneret and Jeannerod 1998. Copyright
© 1998 Elsevier Science Ltd.)
responsible for their actions when these are performed
deliberately. Intentional binding is increased when
associated with outcomes that have moral conse-
quences. It is reduced for actions that have been com-
manded by others, rather than performed freely. These
results do not address the question of whether or not
free will exists, but they suggest that our conscious
experience of acting freely has a major role in creating
social norms of responsibility. Such norms are critical
for maintaining social cohesion.
Unconscious inference occurs in the motor domain
as well as the sensory domain. Our experience of
agency is created from two components: our prior
expectations and the sensory consequences of the
outcome of the action. We are surprised if the actual
sensations do not match what we expect, as when we
pick up an object that is much lighter than anticipated
(Chapter 30). If the outcome confirms our expectations,
however, we pay little attention to the actual sensory
evidence—we experience what we expected to happen
rather than what actually happened.
Pierre Fourneret and Marc Jeannerod asked sub-
jects to draw a vertical line using a computer’s mouse.
The subjects could not see their hand and so could not
see that the computer created a distortion in the line
displayed on the screen. The striking result was that
subjects were not aware that they had moved their
hand at an angle of 10° to the left to produce the vertical
图形
输入板
扰动开始
显示的线
实际
手部动作
10
无意识的调整
计算机
开始点
line on the screen (Figure 59–8). This lack of awareness
occurred for deviations of up to 15°. When subjects
were instructed not to look at the screen but simply
repeat the movement they had just made, they did not
reproduce the deviant movement they had made but
instead drew the straight-ahead movement that they
believed they had made. It would seem that as long as
the goal is realized (drawing a straightforward line),
we experience the expected sensory feedback, not the
actual sensory feedback.
This phenomenon helps us understand some
otherwise bizarre experiences. For example, after the
amputation of a limb, some patients may experience a
phantom limb. They still experience the urge to move
the missing limb, and they can select specific move-
ments they want the missing limb to make. Their
sensorimotor systems predict the proprioceptive sen-
sations they would feel if they were to move an intact
limb, and it is these predicted sensations that underlie
the sensation of a moving phantom limb.
After a limb has been paralyzed due to stroke,
some patients believe that they are still able to move
the limb (anosognosia for hemiplegia). Here, again,
such patients can select the movements they want to
make and are aware of their expectations about the
movement. Despite the lack of sensory evidence that
follows their attempt to initiate the movement, they
believe that the movement did occur.
目标
Kandel-Ch59_1473-1487.indd 1481 18/12/20 11:41 AM
图 59.3.3: 可以在不知不觉中修改动作。受试者被要求用电脑鼠标画一条直线。他们可以看到屏幕上的线条，但
看不到他们的手部动作。计算机被编程为系统地扭曲屏幕上显示的线条。在这里显示的结果中，受试者必须将
他的手向左移动 10°，以在屏幕上产生一条垂直线。目标并不知道做出这样的调整
[539]
。
这种现象有助于我们理解一些原本奇怪的经历。例如，在肢体截肢后，一些患者可能会出现幻肢。他们仍然
有移动缺失肢体的冲动，并且可以选择希望缺失肢体进行的特定动作。他们的感觉运动系统预测了如果他们移
动一个完整肢体时会感受到的本体感觉，正是这些预测的感觉构成了移动幻肢的感觉基础。
在肢体因中风而瘫痪后，一些患者认为他们仍然能够移动肢体（偏瘫的失认症）。同样，在这里，这些患者
可以选择他们想要进行的动作，并意识到他们对动作的期望。尽管在他们尝试开始运动后缺乏感官证据，但他
们相信运动确实发生了。
1312
59.4 有意识地回忆是一个创造性的过程
59.4 有意识地回忆是一个创造性的过程
对于我们大多数人来说，记忆是对过去经历的有意识的想象重温。然而，如果我们不考虑主观经验（行为主
义立场），记忆则是一种使我们过去经历改变未来行为的过程。但在没有对记忆进行有意识回忆或者意识到其对
我们产生影响的情况下。这种类型的经历再次在患有特定脑部损伤的患者中最为显著地出现。
一些患者在颞叶内侧区域受损后变得严重遗忘。根据智商测试，他们的智力没有下降，但无法记住超过几
分钟的事情。尽管这具有破坏性，但这种记忆障碍影响实际上是相当有限的。这个问题主要体现在陈述性记忆
中，最严重的是一种称为情景记忆的陈述性记忆，即回忆生活中事件的能力（第 54 章）。意识在其中起次要作
用（第 53 章）的程序性记忆保持完好。因此，患者仍然可以记住骑自行车等运动技能，并且通常可以正常速度
学习新的运动技能。脑损伤的这种选择性作用会导致剧烈的解离。每周都在学习一些新技能的患者会否认以前
曾执行过这项任务。然后他惊讶地发现自己变得如此熟练。
一项广泛使用的协议测试了受试者回忆他们已经记住的单词列表的能力，这是一项涉及陈述性记忆形式的
任务。在回忆阶段，受试者会看到学习列表中的单词列表以及新单词。遗忘症患者很难完成此类任务，并且可能
会将大部分以前见过的单词错误的归类为新单词，因为她不记得以前见过这些单词。然而，阅读旧词引发的大
脑活动与新词引发的不同：存在一种无意识地认识到差异的认知，相当于单侧忽视或面容失认症患者表现出的
差异。普通受试者通常会发现这项任务很容易，但他们偶尔也会将旧词错误分类为新词；如图 59.4.1 所示，与遗
忘症患者一样，正常受试者的诱发脑反应记录下了被意识回忆所遗失的差异。
1482 Part IX / Diseases of the Nervous System
The Conscious Recall of Memories Is a
Creative Process
For most of us, memory is the conscious imaginative
reliving of a past experience. If we take no account of
subjective experience (the behaviorist stance), however,
memory is a process by which our past experience alters
future behavior. Our behavior is often affected by past
experience, but without conscious recall of the memory
or awareness of the influence it is having on us. Once
again, this type of experience is seen most strikingly in
patients with damage to specific areas of the brain.
Some patients become densely amnesic after dam-
age to the medial regions of the temporal lobe. They
show no decline in intellect as measured by IQ tests
but cannot remember anything for more than a few
minutes. Although devastating, this memory impair-
ment is actually rather circumscribed. The problem
is largely manifested in declarative memory, and most
severely in a type of declarative memory called epi-
sodic memory, the ability to recollect events in one’s life
(Chapter 54). Procedural memory, in which conscious-
ness has a minor role (Chapter 53), remains intact.
Thus, patients can still remember motor skills such as
riding a bicycle and can often learn new motor skills
at a normal rate. This selective effect of brain damage
can lead to dramatic dissociations. A patient who has
been learning some new skill every day for a week will
deny ever having performed the task before. He is then
surprised to find how skillful he has become.
A widely used protocol tests subjects’ ability to
recall lists of words they have memorized, a task that
taps a form of declarative memory. In the recall phase,
a subject is presented with a list of the words that were
on the study list plus new words. An amnesic patient
has great difficulty with this type of task and may
misclassify most of the previously seen words as new
since she cannot recall seeing them before. Neverthe-
less, the brain activity elicited by reading old words is
different from that elicited by the new words: There is
unconscious recognition of a difference, equivalent to
that shown by patients with unilateral neglect or pros-
opagnosia. Normal subjects usually find this task easy,
but they too will occasionally misclassify old words
as new; as with amnesiacs, evoked brain responses
in normal subjects register the distinction lost to con-
scious recall (Figure 59–9).
Occasionally, a subject misclassifies a new word as
an old one. This misclassification amounts to a false
memory. Such misclassifications are most likely to
occur when the new word is semantically related to
one or more of the old words. If the list of old words
contained big, great, huge, then the new word large is
Figure 59–9 Brain activity shows the imprint of forgotten
memories.Subjects were presented with a list of words,
including some that had been presented earlier and some that
were new. When asked to identify the words presented earlier,
subjects correctly identified some of the old words but forgot
others. Immediately after the visual presentation of a word,
there is a brief fluctuation in the evoked potential in the brain.
Evoked responses in the parietal region of the brain reflect
whether or not the words had been seen before, even when
subjects did not consciously recognize the words. The pattern
produced by old words, whether recognized or not, is different
from that produced by the new words. (Reproduced, with permis-
sion, from Rugg et al. 1998. Copyright © 1998 Springer Nature.)
0
新测试项
以前看到
但未识别的测试项
2 微伏
likely to be identified as old. One explanation for this
is that the perception of the new word large has been
unconsciously primed by the previous presentation of
the old words. Thus, the new word large is processed
easily and quickly, and because the subject is aware of
this, he concludes the word must be familiar and clas-
sifies it as old.
This observation emphasizes that memory is a
creative process. Our conscious memories are con-
structed from both conscious recall and unconscious
knowledge. To guard against false memories, as with
false percepts, we use our knowledge about the world
to determine which memories are plausible.
In some patients, the process by which memories
are screened can become dramatically disturbed. If asked
what happened yesterday, most patients with amne-
sia will say that they cannot remember, but a few will
give elaborate accounts that do not correspond to real-
ity. Such false memories are called confabulations and can
sometimes be extremely implausible. For example,
one patient said that he had met Harold Wilson
Kandel-Ch59_1473-1487.indd 1482 18/12/20 11:41 AM
500 毫秒
图 59.4.1: 大脑活动显示被遗忘记忆的印记。向受试者展示了一个单词列表，其中包括一些较早出现的单词和一
些新单词。当被要求识别之前出现的单词时，受试者正确识别了一些旧单词但忘记了其他单词。在一个词的视
觉呈现之后，大脑中的诱发电位会立即出现短暂的波动。大脑顶叶区域的诱发反应反映了这些词是否以前见过，
即使受试者没有有意识地认出这些词。旧词产生的模式，无论是否被识别，都与新词产生的模式不同
[540]
。
有时，受试者会将新词错误分类为旧词。这种错误分类相当于错误的记忆。当新词与一个或多个旧词在语
义上相关时，最有可能发生这种错误分类。如果旧词列表包含 big、great、huge，那么新词 large 很可能被识别为
旧单词。对此的一种解释是，对新词 large 的感知已经在不知不觉中被先前出现的旧词所激发。因此，新词 large
1313
59.5 行为观察需辅以主观报告
的处理变得容易且快速，并且因为受试者意识到了这一点，所以他就会认为这个词一定很熟悉，并将其归类为
旧词。
这一观察强调记忆是一个创造性的过程。我们有意识的记忆由有意识回忆和无意识知识构成。就像防止虚
假感知一样，为了防止虚假记忆，我们利用对世界的了解来确定哪些记忆是可信的。
在某些患者中，筛选记忆的过程会变得非常混乱。如果被问到昨天发生了什么，大多数遗忘症患者会说他
们不记得了，但也存在少数人会详细描述与事实不符的情况。这种错误的记忆被称为虚构，有时可能难以置信。
例如，一名患者说他见过英国前首相哈罗德 · 威尔逊，并讨论了他们正在从事的一项建筑工作。
在想象未来可能发生的事件时，还涉及重建过去情节记忆所需的创造性机制。在海马体受损的遗忘症患者
中，想象新事件的能力明显受损。
59.5 行为观察需辅以主观报告
到 20 世纪中叶，人们已经很清楚经典的行为主义方法显然不足以探索许多心理过程。语言习得、选择性注
意力和工作记忆等过程不能仅根据刺激和反应之间的关系来理解，无论假设的关系有多么复杂。
一些认知过程无意识的证明了我们需要更进一步远离行为主义。如果我们想探索整个意识和无意识认知过
程的范围，仅仅依靠观察外在行为是不够的。我们不能假设一个主体做出有目的、以目标为导向的行为一定会意
识到引发该行为的刺激，甚至是该行为本身。我们必须在行为观察之外补充主观报告。我们需要问受试者：“你
看到刺激了吗？你移动了你的手吗？”
一百年前，内省是心理学中获取数据的主要方法。除此之外，还有什么方法可以研究意识呢？但是不同的心
理学流派得到了不同的结果，而且正如约翰 · 布鲁德斯 · 华生所强调的那样，似乎没有客观的方法来决定谁是正
确的。您如何独立确认主观体验？因此，该方法声名狼藉。在行为主义主导心理学的几十年里，主观报告被认为
是不合适的数据来源。因此，记录主观报告的方法远远落后于记录外显行为的方法。遗憾的是，由于长期以来一
直有排除此类报告的传统，许多认知过程研究仍然不需要受试者的主观体验报告。
继续使用主观报告的一个心理学领域是心理物理学，费希纳于 1860 年引入了对感觉（物理能量）和知觉（心
理体验）之间关系的研究。这些研究给出了稳健且可靠的结果，并创建了一些心理学中为数不多的定律，例如
韦伯定律（2 个刺激之间的可察觉差异与刺激的强度成比例）。在这些研究中，受试者通常会被问到“你看到刺
激了吗？”或“你对自己是否看到了刺激有多大的信心？”
信号检测理论于 1950 年代发展起来，它提供了一种可靠的方法来独立地衡量探测刺激的能力（第 17 章）而
不受任何报告偏见的影响（辨别力，d’）。如果你的辨别能力很高，那么你将成功地检测到刺激的微小变化。最
近，人们对第二个问题越来越感兴趣，“你有多大的把握你看到了刺激？”报告一个人的信心需要元认知能力，即
反思我们的认知过程的能力。这种能力在行为控制中起着重要作用。例如，如果我们意识到自己在某项任务上
表现不佳，我们可能会放慢速度并更加注意我们正在做的事情。
反思我们感知的能力是可以客观衡量的。同样，反思我们认知过程质量的能力也可以得到定量的评估。如
果你的元认知准确性很高，那么你将成功地区分正确和错误的答案。换句话说，正确的检测通常与高置信度相
关联，而不正确的检测与低置信度相关联。但是，你的元认知准确性不一定与您的信号检测能力相关。你可能擅
长检测信号，但同时不擅长判断你的答案是否可能正确或错误。事实上，前额叶皮层受损的患者保留了检测视
觉信号的能力，但在元认知准确性方面表现出明显的缺陷。
口头报告不能在实验动物或不不能说话的婴儿的信号检测实验中使用。一种替代方法是识别反映信心的行
为方面。例如，如果我们确信我们把钥匙落在了客厅的某个地方，我们会在转向走廊之前花更多时间在客厅里
找。路易丝 · 古皮尔和西德 · 欧伊德将这一见解应用于学前婴儿的元认知研究中。婴儿必须记住 2 个盒子中的哪
一个装有玩具，后来在他们不知情的情况下将其拿走了。他们花了更多时间在正确的盒子里搜索。婴儿也更有
可能向成人寻求帮助以打开正确的盒子。这些效应在较长的间隔之后并不会发生。这种行为表明婴儿对自己当
前的知识状态有一定的认识。他们知道在什么时候已经不记得哪个是正确的盒子了。类似的实验表明，老鼠和
猴子也有一些元认知能力。
1314
59.5 行为观察需辅以主观报告
59.5.1 主观报告的验证具有挑战性
主观经验的报告，例如信心，就像一个仪表。正如电表将电阻转换为表盘上指针的位置（读数为 100 欧姆）
一样，受试者将光刺激转换为颜色报告（“我看到红色”）。但是仪表与人的关键之处不同。仪表不会体验红色，
也无法传达其意义。而且，尽管仪表可能有故障，但它永远不会在真正看到蓝色时假装看到红色。大多数时候，
我们假设主观报告是真实的，也就是说，受试者正在尽可能准确地描述他的经历。但是我们如何确定我们可以
相信这些主观报告呢？
使用脑成像可以解决部分验证主观报告的问题。大脑成像研究表明，在与任何明显行为无关的心理活动期
间，大脑的特定区域会出现神经活动。这种心理活动的内容，比如想象或白日梦，只能从受试者的报告中获知。
如果我们扫描一个目标，而他说他正在想象移动他的手，则会在运动系统中的许多部分都能检测到活动。在
大多数运动区域，这种活动的强度比实际运动相关的活动要低，但远高于静息水平。同样地，如图 59.5.1 所示，
如果受试者报告说她正在想象她最近看过的一张脸，则可以在梭状回（“面部识别区域”）中检测到活动。在这
些示例中，扫描仪检测到的神经活动的位置为受试者报告的体验内容提供了独立的确认。在某些有限的情况下，
可以从神经活动的模式推断出意识内容。
1484 Part IX / Diseases of the Nervous System
Figure 59–10 Imagining a face or a place correlates
with activity in specific areas of the brain.Subjects were
scanned while they viewed or imagined faces and houses.
In the first block of trials, subjects alternately viewed a face
or a house. When viewing a face, brain activity increases in
the fusiform face area of the inferior temporal lobe (FFA).
When viewing a house, brain activity increases in the
parahippocampal place area of the inferior temporal cortex
(PPA). In the next block of trials, subjects alternately imagined
a face and a house. The same brain regions are active during
both the imagining and direct viewing of faces and houses,
although the activity is less pronounced during the imagined
viewing. (Reproduced, with permission, from O’Craven and
Kanwisher 2000. Copyright © 2000 MIT.)
the correct box. Similar experiments suggest that rats
and monkeys also have some metacognitive abilities.
Verification of Subjective Reports Is Challenging
Reports of subjective experience, such as confidence,
serve like a meter. Just as an electrical meter converts
electrical resistance into the position of a pointer on a dial
(reading 100 ohms), so a subject converts a light stimulus
into the report of a color (“I see red”). But there is a criti-
cal way in which the meter is not like a person. The meter
does not experience red and cannot communicate mean-
ing. And, although the meter might be faulty, it can never
pretend to see red when it is really seeing blue. Most of
the time, we presume that subjective reports are true,
that is, the subject is trying as far as possible to give an
accurate description of his experience. But how can we be
sure that we can rely on these subjective reports?
查看 想像 查看 想像
海马旁回激活（% 信号改变）
梭状回面孔区激活（% 信号改变）
0
2.5
2
1.5
1
.5
–.5
–1
0
1.5
1
.5
–.5
The problem of verifying subjective reports can
partially be addressed with the use of brain imaging.
Brain imaging studies have shown that neural activ-
ity occurs in localized areas of the brain during mental
activity that is not associated with any overt behavior.
The content of such mental activity, such as imagining
or daydreaming, can be known only from the subject’s
reports.
If we scan a subject while he says he is imagining
moving his hand, activity will be detected in many
parts of the motor system. In most motor regions,
this activity is less intense than the activity associ-
ated with an actual movement, but it is well above
resting levels. Similarly, if a subject reports that she is
imagining a face she has recently seen, activity can be
detected in the fusiform gyrus, the “face recognition
area” (Figure 59–10). In these examples, the location of
the observed neural activity detected by the scanner
Kandel-Ch59_1473-1487.indd 1484 18/12/20 11:41 AM
图 59.5.1: 想象一张脸或一个地方与大脑特定区域的活动相关。受试者在观看或想象面孔和房屋时被扫描。在第
一组试验中，受试者交替观看一张脸或一座房子。观看面部时，下颞叶梭状回面孔区的大脑活动会增加。在看房
子时，下颞皮层海马旁回的大脑活动会增加。在接下来的试验中，受试者交替想象一张脸和一座房子。在想象和
直接观看面孔和房屋时，相同的大脑区域都处于活跃状态，尽管在想象观看时这种活动不太明显
[541]
。
59.5.2 装病和癔症会导致不可靠的主观报告
如果受试者报告说看到了“蓝色”，即使他们体验到的是红色怎么办？怎么会出现这种情况，这种情况下的
主观报告情况如何？
考虑一位因内侧颞叶皮层严重受损而失忆的患者。当给病人看一张他每天在病房里见到的人的照片，尽管
生理测量（脑电图或皮肤电导）显示对这张照片有反应（对他没见过的人的照片没有反应），但病人否认曾经见
1315
59.6 亮点
过这个人）。我们得出的结论是，有意识的记忆过程已受损，而无意识的过程则保持完整。这位患者的主观报告
是他意识到事物的准确描述，但排除了他“知道”的那些没有进入意识的事情。
另一名病人被发现在街上游荡，没有脑损伤的迹象，但报告说他无法记得关于自己或自己历史的任何事情。
当向他展示过去人的照片时，他否认对他们有任何了解，但与此同时，他对照片表现出生理反应。在这种情况
下，由于缺乏可检测的脑损伤（以及记忆丧失的其他特征），我们开始怀疑他陈述的真实性。也许生理反应表明
他在意识中认识到了人。随后，警方确认了患者的身份，我们发现他因在邻近县犯下严重罪行而被通缉。我们
对他的陈述可靠性的怀疑进一步增加了。最后，当他愚蠢地告诉一位病友：“愚弄那些临床心理学家太容易了”，
我们的怀疑得到了证实。
在这种情况下，我们有直接证据表明患者故意误导他人对自己的看法。要欺骗别人，我们不仅要意识到自己
的心理状态，还要意识到他人的心理状态。有什么方法可以测试欺骗吗？一种方法是使用前面讨论的那种记忆
测试。患者学习一组单词。然后向他展示一个新列表，其中包含他刚刚学过的单词和新单词，他必须决定每个单
词是旧单词还是新单词。一个真正的遗忘症患者认不出任何一个词；他将不得不猜测，通过无意识的启动效应，
他会比随机猜测表现得更好。装病的病人可以认出这些旧词，但会他强烈否认以前见过它们。除非他非常老练，
否则他的表现可能比随机猜测还差。我们似乎应该能够区分真正的失忆者和伪装者。
第三种病人也假装失忆（或其他一些疾病），但这样做是无意识的，因此不是装病者。这种情况被称为癔症
性或心因性失忆症。像装病者一样，他在识别测试中的表现比随机测试情况更糟。尽管如此，他并没有意识到他
自己在模拟。同样的机制发生在被催眠的正常人身上，然后被告知他们对刚刚发生的事情没有记忆。这种现象
有时被称为分离状态：大脑中记录经验和进行口头报告的部分已经与创建模拟的部分分离。癔症性的模拟也会
造成感觉丧失，例如癔症性失明和运动障碍，例如癔症性麻痹或癔症性肌张力障碍。
我们距离理解这些疾病的认知过程或潜在生理学还有很长的路要走。一个关键问题是如何区分癔症和伪装。
从意识体验的角度来看，这 2 种障碍是完全不同的：装病者意识到自己在假装，而癔症患者则不然。然而，在这
2 个案例中，患者的主观报告和外显行为非常相似。难道没有可以区分这些不同疾病的措施吗？也许证明这些不
同意识状态之间关键区别的唯一方法是通过神经影像学研究。
59.6 亮点
1. 精神障碍的研究迫使我们面对精神和身体之间的概念鸿沟。现在，不再可能认为精神障碍有精神原因，而
身体障碍有身体原因。
2. 认知神经科学对我们弥合这一差距的尝试产生了重大影响，因为它的描述性语言（即信息处理语言）可
以同时应用于心理过程和神经过程。信息论和计算机的发展暗示了科学怎样解决“主观体验如何从物理大脑的
活动中产生”的问题。
3. 现在很清楚，知觉、行动和记忆是许多并行过程的结果，尽管其中一些过程支持有意识的体验，但大多
数发生在意识水平以下。
4. 当其中一些过程受损而其他过程完好无损时，就会出现显著异常。一名下颞皮层受损的患者 D.F. 不再能
意识到物体的形状，因此无法描述它或识别它是什么。尽管如此，她还是可以将她的手摆成合适的形状来拿起
物体。
5. 我们对自己行为的细节知之甚少，但我们清楚地意识到处于控制之中（行动的控制感）。在极端情况下，
这种行动的控制感可能会与行动控制脱节。肢体截肢后，许多人会感到有一个虚拟的假肢；在肢体因中风而瘫
痪后，一些患者认为他们仍然可以活动肢体。
6. 回忆过去不像重放视频。回忆是一个基于不完美召回并充满常识的创造过程。由于失去了这种创造力，遗
忘症患者很难想象未来，也很难记住过去。
7. 主观体验是人类生活的重要组成部分。当我们做出决定时，我们的选择由我们的行为表明，但我们对该
选择的信心是一种主观体验。我们可以通过口头报告来研究这样的经历。对我们选择的信心是元认知的一个例
子（即反思我们认知过程的能力）。额叶皮层的损伤会损害元认知，但决策能力仍然完好。
1316
59.6 亮点
8. 口头报告并不总是可靠的。人们可以假装失忆以逃避法律制裁。这种装病很难被发现，因为它与癔症性
遗忘等疾病非常相似，在这种情况下，患者并没有意识到自己正在模拟这种疾病。认知神经科学面临的挑战是
区分这些情况。
1317